Study Of Methods For Automated Testing Web Applications Based On The Analysis Of The HTML Code

Due to the dynamic, running complexity and diversity of Web application, there is an urgent need for automated testing methods and tools in order to fully and frequently testing, improving the quality and productivity of Web applications. Current Web automation tools can not handle the form test data generation and Web states effectively. Therefore, it is insufficient for testing and efficiency is low.This paper proposes an algorithm of form pairwise testing based on constraints is used to generate test data for the Web application form testing. This proposed approach not only improves the adequacy of the form testing, but improves the coverage rate of the navigation graph of Web application. Additionally, this paper also proposes a framework for automatic testing methods of Web application based on AJAX technology. We define the Web states as the embedded browser’s DOM tree through the runtime in this framework. A navigation graph of Web application which can guidance the automated testing process and test case set generation is constructed step by step through the depth-first search method. We present an algorithm of Web states similarity to avoid the state space explosion. Page noise and similar elements in Web states are also handled to avoid searching unnecessary paths.Based on previous works, we present our automation testing tool called ComjaxTest, by which empirical study is conducted on eight Web applications. The experimental results show that, the coverage rate of nodes and edges of the navigation graph of Web applications can reach more than 79.68% and 70.29% respectively, higher than existing tools Crawljax of 66.31% and 64.16%. Our constraint-based combination method covers 100% form submission results which is far higher than Crawljax of 57.89%. By this approach we can improves the coverage rate of the navigation graph and the ability of detecting errors in ComjaxTest.